Open-source observability platform for AI agents that catches failures, surfaces root causes, and confirms fixes.
Laminar is an open-source observability and debugging platform for AI agents that provides end-to-end failure detection via plain-language Signals with Slack alerts, readable transcripts and timelines of agent runs, and root-cause analysis down to the exact failing step. The platform clusters recurring issues, tracks them over time, auto-resolves fixed clusters, and re-opens them if issues return. It includes a CLI and MCP integration for coding agents to iterate on fixes, screen recording for browser agents, full SQL access to platform data, and 20x more storage-efficient ingestion by storing only unique content. Laminar is Apache 2.0 licensed, self-hostable via Docker or Helm on AWS/GCP, and SOC 2 Type II and HIPAA compliant. Laminar is a product of Laminar.
For people
For agents
Nothing listed yet.